Predictive Dynamic System Scheduling

• US 20110247003A1
• Filed: 03/31/2010
• Published: 10/06/2011
• Est. Priority Date: 03/31/2010
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Seelam, Seetharami, Salapura, Valentina, Fong, Liana L.
• Granted Patent: US 8,806,501 B2
• Time in Patent Office: Days
• US Class Current: 718/104
• CPC Class Codes: G06F 9/5061 Partitioning or combining of resources

Abstract

Resources of a partitionable computer system are partitioned into at least first and second partitions, in accordance with a first or second mode of operation of the partitionable computer system. The system is run in the first or second mode, partitioned in accordance with the partitioning step. Periodically, it is determined whether the computer system should be switched from one mode to the other mode. If so, the computer system is run in the other mode, partitioned in accordance with the other mode. The first and second modes of operation are defined in accordance with historical observations of the partitionable computer system. The periodic determination is carried out based on predictions in accordance with the historical observations.

Claims

1. A method comprising:

partitioning resources of a partitionable computer system into at least first and second partitions, in accordance with one of a first mode of operation of said partitionable computer system and a second mode of operation of said partitionable computer system, said first and second modes of operation being defined in accordance with historical observations of said partitionable computer system;

running said partitionable computer system, in said one of said first and second modes, partitioned in accordance with said partitioning step;

periodically determining whether said computer system should be switched from said one of said first and second modes to another of said first and second modes, said periodic determination being carried out based on predictions in accordance with said historical observations; and

responsive to said periodic determination being affirmative, running said computer system, in said another of said first and second modes, partitioned in accordance with said another of said first and second modes.

2. The method of claim 1, further comprising:

setting at least one first mode threshold for said first mode of operation of said partitionable computer system and at least one second mode threshold for said second mode of operation of said partitionable computer system;

while operating in a given one of said modes, periodically re-evaluating said partitioning against;

if running in said first mode, said at least one first mode threshold; and

if running in said second mode, said at least one second mode threshold; and

if said periodic re-evaluation suggests that one of said first and second partitions for said one of said first and second modes is underutilized, dynamically re-partitioning said resources of said partitionable computer system to reassign at least some of said resources of said partitionable computer system from said underutilized one of said first and second partitions for said one of said first and second modes to another one of said first and second partitions for said one of said first and second modes.

3. The method of claim 2 wherein said re-evaluating step comprises comparing at least one of a number of jobs in said first partition and a number of jobs in said second partition against:

said at least one first mode threshold, if running in said first mode; and

said at least one second mode threshold, if running in said second mode.

4. The method of claim 3, wherein, in said comparing:

said at least one first mode threshold comprises a high water mark threshold for said at least one of said number of said jobs in said first partition and said number of said jobs in said second partition, when running in said first mode; and

said at least one second mode threshold comprises a high water mark threshold for said at least one of said number of said jobs in said first partition and said number of said jobs in said second partition, when running in said second mode.

5. The method of claim 2, further comprising:

while running said computer system in said another of said first and second modes, periodically determining whether said computer system should be switched from said another of said first and second modes back to said one of said first and second modes, said periodic determination being carried out based on predictions in accordance with said historical observations; and

responsive to said periodic determination, during said running in said another of said first and second modes, being affirmative, running said computer system, in said one of said first and second modes, partitioned for said one of said first and second modes.

6. The method of claim 2, wherein said periodic re-evaluation step is carried out upon job termination.

7. The method of claim 2, wherein said periodic re-evaluation step is carried out upon reaching a job checkpoint.

8. The method of claim 2, wherein said periodic re-evaluation step is carried out upon passage of a predetermined time interval.

9. The method of claim 2, further comprising developing, based on said historical observations:

said at least one first mode threshold; and

said at least one second mode threshold.

10. The method of claim 2, further comprising periodically running said partitionable computer system in a third mode comprising a maintenance mode which is not partitioned.

11. The method of claim 2, further comprising providing a system, wherein the system comprises distinct software modules, each of the distinct software modules being embodied on a computer-readable storage medium, and wherein the distinct software modules comprise a partition tracking module, a re-partition calculating module, a mode-switching module, and learning module;

wherein;

said partitioning of said resources into said at least first and second partitions, in accordance with said one of said first mode of operation of said partitionable computer system and said second mode of operation of said partitionable computer system, is carried out by said partition tracking module executing on at least one hardware processor;

said periodically determining whether said computer system should be switched from said one of said first and second modes to another of said first and second modes, is carried out by said mode-switching module executing on said at least one hardware processor;

said setting of said at least one first mode threshold for said first mode of operation of said partitionable computer system and said at least one second mode threshold for said second mode of operation of said partitionable computer system, is carried out by said learning module executing on said at least one hardware processor; and

said periodically re-evaluating of said partitioning is carried out by said re-partition calculating module executing on said at least one hardware processor.

12. A computer program product comprising a computer readable storage medium having computer readable program code embodied therewith, said computer readable program code comprising:

computer readable program code configured to partition resources of a partitionable computer system into at least first and second partitions, in accordance with one of a first mode of operation of said partitionable computer system and a second mode of operation of said partitionable computer system, said first and second modes of operation being defined in accordance with historical observations of said partitionable computer system;

computer readable program code configured to run said partitionable computer system, in said one of said first and second modes, partitioned in accordance with said partitioning step;

computer readable program code configured to periodically determine whether said computer system should be switched from said one of said first and second modes to another of said first and second modes, said periodic determination being carried out based on predictions in accordance with said historical observations; and

computer readable program code configured to, responsive to said periodic determination being affirmative, run said computer system, in said another of said first and second modes, partitioned in accordance with said another of said first and second modes.

13. The computer program product of claim 12, further comprising:

computer readable program code configured to set at least one first mode threshold for said first mode of operation of said partitionable computer system and at least one second mode threshold for said second mode of operation of said partitionable computer system;

computer readable program code configured to, while operating in a given one of said modes, periodically re-evaluate said partitioning against;

if running in said first mode, said at least one first mode threshold; and

if running in said second mode, said at least one second mode threshold; and

computer readable program code configured to, if said periodic re-evaluation suggests that one of said first and second partitions for said one of said first and second modes is underutilized, dynamically re-partition said resources of said partitionable computer system to reassign at least some of said resources of said partitionable computer system from said underutilized one of said first and second partitions for said one of said first and second modes to another one of said first and second partitions for said one of said first and second modes.

14. The computer program product of claim 13 wherein said computer readable program code configured to re-evaluate comprises computer readable program code configured to compare at least one of a number of jobs in said first partition and a number of jobs in said second partition against:

said at least one first mode threshold, if running in said first mode; and

said at least one second mode threshold, if running in said second mode.

15. The computer program product of claim 14, wherein, in said computer readable program code configured to compare:

said at least one first mode threshold comprises a high water mark threshold for said at least one of said number of said jobs in said first partition and said number of said jobs in said second partition, when running in said first mode; and

said at least one second mode threshold comprises a high water mark threshold for said at least one of said number of said jobs in said first partition and said number of said jobs in said second partition, when running in said second mode.

16. The computer program product of claim 13, further comprising:

computer readable program code configured to, while running said computer system in said another of said first and second modes, periodically determine whether said computer system should be switched from said another of said first and second modes back to said one of said first and second modes, said periodic determination being carried out based on predictions in accordance with said historical observations; and

computer readable program code configured to, responsive to said periodic determination, during said running in said another of said first and second modes, being affirmative, running said computer system, in said one of said first and second modes, partitioned for said one of said first and second modes.

17. The computer program product of claim 13, wherein said computer readable program code configured to periodically re-evaluate is configured to carry said periodic re-evaluation out upon job termination.

18. The computer program product of claim 13, wherein said computer readable program code configured to periodically re-evaluate is configured to carry said periodic re-evaluation out upon reaching a job checkpoint.

19. The computer program product of claim 13, wherein said computer readable program code configured to periodically re-evaluate is configured to carry said periodic re-evaluation out upon passage of a predetermined time interval.

20. The computer program product of claim 13, further comprising computer readable program code configured to develop, based on said historical observations:

said at least one first mode threshold; and

said at least one second mode threshold.

21. The computer program product of claim 13, further comprising computer readable program code configured to periodically run said partitionable computer system in a third mode comprising a maintenance mode which is not partitioned.

22. An apparatus comprising:

a memory; and

at least one processor, coupled to said memory, and operative to;

partition resources of a partitionable computer system into at least first and second partitions, in accordance with one of a first mode of operation of said partitionable computer system and a second mode of operation of said partitionable computer system, said first and second modes of operation being defined in accordance with historical observations of said partitionable computer system;

run said partitionable computer system, in said one of said first and second modes, partitioned in accordance with said partitioning step;

periodically determine whether said computer system should be switched from said one of said first and second modes to another of said first and second modes, said periodic determination being carried out based on predictions in accordance with said historical observations; and

responsive to said periodic determination being affirmative, run said computer system, in said another of said first and second modes, partitioned in accordance with said another of said first and second modes.

23. The apparatus of claim 22, wherein said at least one processor is further operative to:

set at least one first mode threshold for said first mode of operation of said partitionable computer system and at least one second mode threshold for said second mode of operation of said partitionable computer system;

while operating in a given one of said modes, periodically re-evaluate said partitioning against;

if running in said first mode, said at least one first mode threshold; and

if running in said second mode, said at least one second mode threshold; and

if said periodic re-evaluation suggests that one of said first and second partitions for said one of said first and second modes is underutilized, dynamically re-partition said resources of said partitionable computer system to reassign at least some of said resources of said partitionable computer system from said underutilized one of said first and second partitions for said one of said first and second modes to another one of said first and second partitions for said one of said first and second modes.

24. The apparatus of claim 23, further comprising a plurality of distinct software modules, each of the distinct software modules being embodied on a computer-readable storage medium, and wherein the distinct software modules comprise a partition tracking module, a re-partition calculating module, a mode-switching module, and a learning module;

wherein;

said at least one processor is operative to partitioning said resources into said at least first and second partitions, in accordance with said one of said first mode of operation of said partitionable computer system and said second mode of operation of said partitionable computer system, by executing said partition tracking module;

said at least one processor is operative to periodically determine whether said computer system should be switched from said one of said first and second modes to another of said first and second modes, by executing said mode-switching module;

said at least one processor is operative to set said at least one first mode threshold for said first mode of operation of said partitionable computer system and said at least one second mode threshold for said second mode of operation of said partitionable computer system, by executing said learning module; and

said at least one processor is operative to periodically re-evaluate said partitioning by executing said re-partition calculating module.

25. An apparatus comprising:

means for partitioning resources of a partitionable computer system into at least first and second partitions, in accordance with one of a first mode of operation of said partitionable computer system and a second mode of operation of said partitionable computer system, said first and second modes of operation being defined in accordance with historical observations of said partitionable computer system;

means for running said partitionable computer system, in said one of said first and second modes, partitioned in accordance with said partitioning step;

means for periodically determining whether said computer system should be switched from said one of said first and second modes to another of said first and second modes; and

means for, responsive to said periodic determination being affirmative, running said computer system, in said another of said first and second modes, partitioned in accordance with said another of said first and second modes.